The Rewards System

Fairly rewarding contributors via subject and objective data streams.

Take note that the Rewards System is currently in development and is scheduled to be completed during Q4 of 2021.

Rewards System Process

The purpose of this system is to create a way to record, reward and analyze fairly the work done by contributors to the TEC, creating a manner of providing decentralized updates of the action happening in the Commons. By this system we can gather realtime acknowledgement of contributions, in a decentralized manner. We have multiple systems that are designed to reduce as much as possible the mental overhead and time commitment of quantification and also average out the subjectivity of praise in itself, by having a decentralized oracle of quantifiers. This is our current proposed process by which we will achieve the goals of this system.

Step 1 - Collect Data

The most important step is to collect the data related to the work done by TEC Contributors. This is broken down into subcategories and further into specific input streams. Data is collected continuously and submitted for Quantification on a bi-weekly schedule. The two sub-categories are Automatically Quantified and Manually Quantified. The corresponding inputs for each category are as such:

  1. Auto Quantified a. Github b. Discourse c. Discord Meeting Attendance d. Twitter

  2. Manually Quantified a. Discord Praise b. Telegram Praise

It’s important to note that in general, auto quantification is related to objective contributions and manual quantification to subjective contributions. It’s possible that some actions in Github and Discourse for example fall under the Praise process if they have a subjective character.

Step 2 - Quantification

We aggregate all input streams and run each through their distinct quantification paths.

Automatic Quantification

For Automatic Quantification we will be using a SourceCred instance and bot that will track and quantify specified actions. The Reward Board will have the ability to establish and propose to modify the actions we value and how much they are relatively worth. These “weight” configurations will be shown in the output of Step 7, for every quant. For each input we outline the actions that can be automatically tracked and quantified, the four main sources for automatic quantification are Github, Discourse, Discord Meeting Attendance (Alexandra?), and Twitter. A comprehensive list of the parameters for each source can be found in the Appendix at the bottom of this document.

Manual Quantification

Praise will be recorded and saved to the Praise backend by TEC tailored bots in both Telegram and Discord. The amassed praise will have to be quantified each cycle manually. Manual Quantification will be done by members of the community. We will have a pool of Quantifiers that have opted in for the responsibility of quantifying praise and we will draft a group of 8-10 randomly selected Quantifiers from this pool, bi-weekly, to complete this task. Quantifiers will also be rewarded for the work they do in manual quantification to incentivize participation in the system. The UI and user flow for Quantifiers that will be built will adhere to a few important points:

  • Quantifiers will have three (3) days to asynchronously quantify their assigned praise

  • Praise will be grouped by member being praised inside the UI for quantifiers

  • Praisors and Praisees (Givers and Receivers) will have anonymous aliases to allow Quantifiers to have context but avoid personal subjectivity. As an example:

    • Unicorn praises Beaver for ...

    • Unicorn praises Goldfish for ...

    • Beaver praises Unicorn for …

  • Quantifiers will have overlapping sections of praise that they will need to quantify, not the entire praise for the whole cycle. Meaning that more than 1 quantifier will evaluate each praise and the average will be taken and passed on as the actual quantification

  • A Fibonacci sequence will be used as the incremental factor for quantifying. To the Quantifier this will show up as simply a slider that returns a numerical value.-> more impactful. The range of values go from the least relevant which is 0 to most impactful which is 134, according to this sequence:

    • 0 -> 1 -> 1 -> 2 -> 3 -> 5 -> 8 -> 13 -> 21 -> 34 -> 55 -> 89 -> 144 of which we can reduce to 5 possible values: 0 -> 13 -> 21 -> 55 -> 144

  • Quantifiers will be provided a “Rules of Praise” (WIP) document which will help them navigate the quantification process.

Both streams will merge once they have been quantified in their respective manners and passed onto the next step.

Step 3 - Human Verification

The Quantifiers from this round will verify the token distribution, checking the amounts calculated and flagging any issues or anomalies. Any problems within the calculation algorithm can be caught and corrected or any member trying to game the system can be identified. At the end of the 3 day window for quantifying praise a call will be held for the Quantifiers who wish to review the entire praise sheet to collaboratively analyze and suggest adjustments to each other’s work. This will bring us cultural insights and help us find problems and strengths in the early stages. A template for discussion and analysis will be drafted and provided by the rewards DAO.

Step 4 - Calculate Rewards

Once we have the combined praise data from all streams we’ll use the associated quantification values to calculate the actual token amounts that will be distributed to each eligible recipient. The initial allocation distribution will be 50% of the allocation value to sourceCred and the other 50% goes to praise. We will value both equally in the initial configuration. A percentage of the Common Pool will be requested from the TEC treasury for funding Rewards. This percentage will vary between 0.3% - 1.5%. The Reward Board can modify the source ratio and the rewards allocation percentage, but they must be approved by a community vote.

Step 5 - Final Approval

The final token distribution would then be withdrawn from the TEC treasury and sent to the Rewards DAO (an Aragon DAO deployment). This DAO will be stewarded by a board of 3-7 trusted members (Reward Board) who have been vested with the responsibility of inspecting the final distribution and pushing the button to release funds. They will need to check for any oversights or collusion between Quantifiers. The Vote for releasing rewards DAO funds must meet a Quorum of 41%(at least 3 of 7 must vote) and have minimum 81% Support (2 members voting No can block the proposal).

Choosing the Reward Board

We will have a formal nomination process for the initial Reward Board with up to a maximum of 7 seats. A Reward Board member who has not participated for 3 consecutive rounds will be asked to give up their seat.

Powers of the Reward Board

This DAO will have extraordinary powers over the Rewards System, the greater community however will act as the arbiter in case of metagovernance and distribution modifications. We will use Snapshot for these instances. An exhaustive list of this board’s powers are as follows:

  1. Change the weight configurations in SourceCred

  2. Adjust the final token distribution for each contributor in each rewards round

  3. Change the allocation percentage for each round of rewards (the TEC treasury still must approve sending the funds).

  4. Modify the distribution ratio between SourceCred and Praise

  5. Mint tokens for new board members, burn tokens for outgoing members

  6. Modify the distribution percentages of the three entities outlined in Step 6

...But Who Watches the Watchers?

The community will have several instances to act as the backstop in instances of collusion or poor judgement from board members. We define checkpoints for each power numbered in the previous sub-section:

  • Action 1 will be output in Step 7 allowing the community to see the weights used and flag issues.

  • Action 2 will be recorded using a Git styled approach where modifications to the distribution will be saved as “commits” with details such as who made modifications, when, and what was modified.

  • Actions 3-6 will require community votes to approve proposals made by the board.

Step 6 - Distribution

Reward Allocations are valued in DAI, but paid out in TEC. We want to inject our contributors directly into our token economy and we’ll use our own Augmented Bonding Curve to achieve that. The ABC will convert the received funds from the TEC treasury, swapping wxDAI to TEC. Rewards DAO funding requests from the TEC treasury will be the Allocation Percentage(% of Common Pool) + Entry Tribute(%) to cover the costs of swapping. When the funds have been released by the Reward Board they will be paid out to three entities - the Reward Board, Quantifiers, and Contributors. The distribution will have already been set by the Reward Board and ratified by the greater community at the initialization of the Rewards System. To aid in visualization we can use this imagined example:

The allocation percentage is 1% of the Common Pool for each rewards round

If the Common Pool has 600,000 wxDAI and the Entry Tribute is 10%, we request 6600 wxDAI

The Price of TEC is currently 2 wxDAI, converting it on the ABC we receive 2887 TEC accounting for price slippage. (rough estimation)

The Reward Board establishes an allocation distribution of 88% to contributors, 10% to Quantifiers, 2% to Reward Board

From these amounts, assuming no modifications were made by the Rewards DAO in Step 5, the token distribution amounts would break down as follows:

  • Contributors receive 2540.56 TEC

    • 1270.28 for SourceCred

    • 1270.28 for Praise

    • Or averaged out across 50 unique contributors renders 50.8 TEC each

  • Quantifiers receives 288.7 TEC

    • If there’s 10 Quantifiers that’s 28.7 TEC each.

  • Reward Board receives 57.74 TEC

    • If there’s 7 board members that’s 8.24 TEC each

Once the amounts are calculated and approved by the DAO vote they are distributed using the Aragon DAO Agent.

Step 7 - Analysis

Run the data analysis of each round according to established metrics. Have a cultural session to integrate insights and inform the parameters of sourcecred and the reach of praise.

A forum post will be generated for each quant showing who received how much funds and the SourceCred weight configurations. By this process we invite analysis and discussions and empower transparency. Data points and more metrics will be added in Post MVP that will allow robust analytics of praise. *A note on record-keeping: Ultimately the blockchain is the final source of truth for observing distributions. However, a ledger of final praise distributions will be maintained and made publicly available, this will be the responsibility of the Reward Board.

Quantification Parameters Appendix

*This list is subject to change as we discover new technical possibilities and limitations

Data Source

Inputs

Parameters and Actions to Weigh

SourceCred

Discourse

  • Like received

  • Like given

  • Topic made

  • Post made

  • Replied to post

  • Post receives reply

  • mention user

  • user is mentioned

  • topic is referenced

  • reference topic

GitHub

  • Create repository

  • Create Issue

  • Create PR

  • Review PR

  • Add Comment

  • Add Commit

  • Give Reaction :+1:

  • Receive Reaction :+1:

  • Give Reaction :heart:

  • Receive Reaction :heart:

  • Give Reaction :rocket:

  • Receive Reaction :rocket:

  • Merge PR

Twitter

  • Retweeted TEC Post

  • Post was retweeted by TEC

  • Mentioned by TEC

  • Mentions TEC

  • uses hashtag #????

Discord Meeting Attendance

  • member attended meeting {x}

  • time spent in meeting?

  • time spent talking in meeting?

Praise

Discord

  • !praise in the TEC Discord

Telegram

  • !praise in TEC Telegram group \

Reward Board

  • Board registers initial rewards parameters config

  • Board approves rewards distribution

Quantifiers

  • Quantifies praise

Contributors

  • Receives quants/praise for work/contributions

Last updated